Munich Personal RePEc Archive

An analysis of mergers and acquisitions

in the Turkish banking sector

Mumcu, Ayşe and Zenginobuz, Unal

2 January 2002

Online at https://mpra.ub.uni-muenchen.de/191/

MPRA Paper No. 191, posted 07 Oct 2006 UTC

Forthcoming in: N. Colton and S. Neaime (Eds.), Money and Finance in the Middle East :

Missed Opportunities or Future Prospects? - Research in the Middle East Economics, Volume

6, 137-167, Elsevier, 2005 (forthcoming).

AN ANALYSIS OF MERGERS AND ACQUISITIONS IN

THE TURKISH BANKING SECTOR

Ayşe Mumcu and E. Ünal Zenginobuz

Department of Economics and Center for Economics and Econometrics Boğaziçi University

80815 Bebek, İstanbul TURKEY

Revised: January 3, 2002

ABSTRACT

We explore various aspects of mergers and acquisitions in the banking industry within a simple

model that allows explicit comparison of sector performance before and after the mergers and

acquisitions. The industry structure we look at involves a few dominant banks and a competitive

fringe, which we take it as the structure most likely to resemble the Turkish banking industry in

the aftermath of the ongoing restructuring process. Using a reasonable set of parameters to

simulate the model, we do comparative statics exercises regarding the impact of mergers among

domestic as well as with foreign banks on equilibrium outcomes.

JEL Classification Numbers

L13, G21, G34

Keywords

banking, mergers and acquisitions, imperfect competition, Turkey

2

1. Introduction

Turkish economy has been hit by two very severe crises recently, namely those of November

2000 and February 2001. These crises caught the Turkish banking sector in the middle of a

sweeping restructuring process, which was one of the critical components of the comprehensive

disinflation program Turkey adopted at the beginning of 2000. The disinflation program

involved tight fiscal and monetary policies, large-scale structural reforms, and a pre-determined

exchange rate policy to serve as a nominal anchor in reducing inflation from its chronically high

levels. Regarding the banking sector, the disinflation program of 2000 foresaw revamping of the

legal and regulatory framework for banking supervision in accordance with EU and world

standards, correcting the weaknesses in the private banking system, and restructuring and the

ultimately privatization of the state banks. Towards this end, the powers of the independent

Banking Regulatory and Supervisory Agency (BRSA), which was established as part of a new

banking law in June 1999, were further strengthened through a series of amendments approved

in the Parliament in December 1999. As the single regulatory and supervisory agency to oversee

the sector, BRSA has completely independent jurisdiction over the entry and exit of banks and

over changes to the regulatory framework.

Prior to the disinflation program the Turkish banking industry had faced distorted

incentives in the chronically high and erratic inflationary environment, which went together with

increasing government deficits over the last decade. Excessive and persistent public sector

borrowing requirements led to very high real returns in government issued securities that

allowed some private banks to accumulate asset portfolios that were far from sound. Together

with the slackening of entry requirements to the sector and the overall weakness of the

regulatory framework, this environment contributed to the fragmentation of the banking sector

into small banks. A significant number of small banks carrying weak asset portfolios became

3

insolvent over time and had to be transferred to the Savings Deposit Insurance Fund (SDIF).1

The first banking crisis in December 2000, as well as the one that followed the financial crisis in

February 2001, can be seen as a culmination of persistent distortions in the Turkish banking

sector.

The new and tougher regulatory framework, together with the new macroeconomic

environment that was to arise from the disinflation program of 2000, was expected to lead to

significant changes in the Turkish banking sector. This included the increased likelihood of

significant mergers and consolidation in the sector. Such a process was also expected to increase

the presence of foreign banks in Turkey through acquisitions. Both of these developments have

gained increased momentum after the already ailing Turkish banking sector was further hit by

the November 2000 and the February 2001 crises.

Understanding the various impacts of mergers and acquisitions requires modeling of the

banking industry structure and behavior of banks. The objective of this paper is to explore

various aspects of mergers and acquisitions within a simple albeit rigorous model that allows

explicit comparison of sector performance before and after mergers and acquisitions. The model

studies strategic interaction among commercial banks in an imperfectly competitive banking in-

dustry, and allows us to consider the determinants of feasibility and desirability of mergers and

acquisitions among domestic as well as foreign banks.

Given the aims of the ongoing fundamental reform process in the Turkish banking

sector, we abstract from much of the features that are currently relevant and concentrate on a

sector structure that is expected to arise after the proposed restructuring is successfully

completed. This amounts to looking at a banking industry in a stabilized economy in which all

public banks are privatized, banking regulations are in place and fully enforced.

1 SDIF was established in 1983. The Turkish Central Bank had managed SDIF from its inception until August 30, 2000, when it was transferred to BRSA on August 30, 2000.

4

The main features of the stylized model we develop in this paper can be summarized as

follows: Adopting an Industrial Organization approach, we consider an imperfectly competitive

banking industry. Banks are assumed to act as independent entities that react strategically to

their environment with deposits received, loan amounts serviced, and net foreign position

viewed as strategic variables. The industry structure we look at involves a few dominant banks

with a (competitive) fringe, which we take as the structure most likely to resemble the Turkish

banking industry in the aftermath of the ongoing restructuring process. For a given as well as

changing number of firms, we first study the features of equilibrium outcomes such as the size

of banking industry, interest rates (deposit, loan, and interbank rates), etc. Using a reasonable

set of parameters to simulate the model, we go on to do various comparative statics exercises

regarding the impact of mergers among domestic as well as with foreign banks on equilibrium

outcomes. Other comparative statics exercises that we carry out include the impact of deposit

insurance, reserve requirement ratios, the country's international credit rating, and the global

economic conditions.

The plan of the paper is as follows: Section 2 provides a quick glance at the Turkish

banking industry structure and summarizes the developments that have rendered it vulnerable to

crises of November 2000 and February 2001. In Section 3 we present the model we study.

Section 4 presents the empirical application. Discussion and conclusions follow in Section 5.

2. A Quick Glance at The Turkish Banking Industry Structure2

At the end of September 2001, there were 68 banks operating in Turkey.3 As of that date, the

total assets in the Turkish banking sector amounted USD 110.7 billion. The deposits stood at

USD 68.8 billion for the whole industry, while the total amount of loans extended by the

2 The information and data presented in this section is compiled from various periodic reports by the Banks Association of Turkey and BRSA. The websites (in English) of the Banks Association of Turkey and BRSA can be reached at http://www.tbb.org.tr/english and http://www.bddk.org.tr/english/mainpage/index_eng.htm, respectively. 3 The most recent official data on the Turkish banking industry is as of September 30, 2001.

5

industry equaled USD 28.6 billion. Further data on the Turkish banking sector as of September

2001 is presented below in Table 1.

TABLE 1 ABOUT HERE

Of the total of 68 banks, 52 are commercial (deposit money) banks while the rest are

investment and development banks. Of the 52 commercial banks, 3 are state banks, 24 are

privately owned domestic banks, 16 are privately owned foreign banks, and 9 banks are under

the management of SDIF.

As of September 2001, the total assets of 24 private domestic banks constituted 53.6% of

the total assets of the Turkish banking sector. The share of state banks in total assets stood at

27.2%, while foreign banks’ share in total assets were 2.6%.

A cursory look at the evolution of Turkish banking sector from early 1980s until the

November 2000 crisis reveals that it has significantly expanded during this period. The number

of banks increased from 43 in 1980 to 66 in 1990 and to 79 by the end of 2000. As noted above,

in the aftermath of the two recent financial crises this number now stands at 68. Total

employment in the banking sector has also increased considerably during the 1980-2000 period:

starting from 125,000 in 1980, it went up to 154,000 in 1990, and reached 170,000 in 2000.

However, since the end of 2000 the number of employees has decreased by almost 15% to

147,453 in the first 9 months of 2001, indicating the extent of the shock received by the

industry.4

Table 2 displays the changes in various financial indicators for the Turkish banking

sector for the 1980-2000 period. Total assets of the Turkish banking sector increased from USD

20.8 billion (28.6% of GNP) in 1980 to USD 58.2 billion (38.2% of GNP) in 1990 and to USD

4 Information on the developments in the Turkish banking sector over the 1980-2000 periods is compiled from various BRSA reports, in particular from BRSA (2001).

6

155 billion (76.9% of GNP) in 2000. The share of state banks in banking sector total assets has

decreased from 45% in 1980 to 34% in 2000. Recall that further restructuring in the first 9

months of 2001 has reduced this figure down to 27.4%. The number of foreign banks increased

from 4 to 18 in 2000. The foreign banks’ share in total banking sector assets increased from

3.7% in 1992 to 5.4% in 2000, but this figure stood at 2.6% at the end of September 2001.

Foreign presence in the Turkish banking sector is thus very low compared to similar emerging

market economies.

TABLE 2 ABOUT HERE

Tables 1 and 2 also reveal that the share of loans in total assets of the banking sector

declined from 47% in 1990 to 33% in 2000, and finally to 28.5% at the end of September 2001.

The ratio of loans to deposits also decreased from 84% in 1990 to 51% in 2000, and to 42.8% as

of September 2001. The credit to GNP ratio stood at a comparatively very low level of 25% at

the end of 2000.

There are a number of reasons why the Turkish banking sector has continually moved

away from its traditional financial intermediation since the beginning of 1990s. The last decade

was one of persistently high and volatile inflation as well as highly erratic growth rates for the

Turkish economy. Liberalization of the current account in 1989 led to external capital inflows,

but these were mostly very short-term in nature. External capital inflows have proven to be very

sensitive to macroeconomic instabilities that distressed the Turkish economy and their fragility,

in turn, further contributed to the volatility of economic variables. This was accompanied by

currency substitution as confidence in Turkish Lira was continuously eroding in the presence of

high inflation. The maturity of bank funding sources has shortened significantly and at the same

time the share of foreign currency liabilities in total liabilities has increased considerably.

7

Another factor that contributed to the increase in foreign exchange open position of the banking

sector was the attractiveness of borrowing abroad to purchase government securities that offered

very high real interest rates. All throughout the 1990s persistent and increasing public sector

borrowing requirements were financed through various government securities sold in the

domestic financial markets.

As for the state banks, their financial health was severely compromised as they were

increasingly being used to finance various government expenditures. This, together with

inefficient management due to appointments based on political affiliation rather than merit, has

resulted in the accumulation of excessive amounts of the so called ‘‘duty losses’’, which led the

illiquid state banks to borrow from the markets at very high rates and short maturities. In turn,

this has contributed to private banks’ practice of borrowing abroad to lend at very high domestic

interest rates.

For all practical purposes, there were no effective internal risk assessment and

management systems that were in use while liquidity, interest rate, and foreign exchange risks in

the Turkish banking sector continually increased throughout the 1990s. In addition, the banking

system was not subject to effective supervision and regulation by an independent public agency.

The increasingly more precarious Turkish banking sector was put through an early test in

the wake of the 1994 financial crisis, which was only contained when interest rates were let to

rise very sharply and the Turkish Lira depreciated by more than 50%. The banks with short

positions in foreign exchange and mismatch of maturities incurred large losses. In addition, a

significant run on deposits occurred which put several banks in severe liquidity problems.

Further deposit withdrawals ensued when three small banks had to be put on the liquidation

block. The run on the banks could only be contained when government introduced hundred

percent guarantee for savings deposits and declared that it would provide liquidity to the banks

in need. These developments revealed that the Turkish banking sector was quite segmented in

8

that it contained a small group of rather efficient and profitable banks and quite a few small

banks at the margin that were prone to falling insolvent in the face of financial downturns.

The Turkish banking sector did recover from the 1994 crisis and managed to grow after

1995. However, a number of key issues, such as risk management and prudent regulation and

supervision of the banking sector, were never seriously addressed. In fact, the hundred percent

guarantee on savings deposits still continue today.

Before the structural reforms that the disinflation program that Turkey adopted at the

beginning of 2000 could be fully implemented and take effect in the area of banking, other

aspects of the program had started to seriously compromise the financial health of the banking

sector once again. The disinflation program initially led to a sharp decline in the interest rates,

but the government’s financing needs continued. The pre-announced exchange rate path and the

real appreciation of the Turkish Lira allowed banks once more to borrow abroad at low rates to

invest in longer term fixed rate Treasury bills. Soaring maturity mismatches and foreign

currency open positions were again the end result.

Delays in structural reforms in several key areas, together with the sharp increase in

current account deficit, led to loss of confidence in the disinflation program by the Fall of 2000.

An outflow of foreign funds ensued, leading to sharp increases in Treasury bill rates, which in

turn led to serious financing difficulties for several banks that had concomitant maturity

mismatch and foreign exchange open positions. In the resulting crisis in November 2000 several

private banks had to be put under the management of SDIF. Despite efforts to calm the markets,

the November 2000 crisis was never really contained, leading to increasing loss of credibility for

the disinflation program, and finally the exchange rate peg, the vital component of the

disinflation program, had to be abandoned in February 2001. In the aftermath of the crisis, once

again a number of banks became insolvent and had to be taken under the umbrella of SDIF.

After the February 2001 crisis a new economic program, named ‘Transition to a Strong

Economy’, was adopted. This program called for completion of the structural transformation of

9

the banking sector and it is still under effect. In addition to efforts to further improve banking

regulation and supervision, state banks and banks under the umbrella of SDIF are currently

undergoing drastic financial restructuring. As Table 1 indicates, the Turkish banking industry

has already undergone a considerable amount of consolidation. There have already been

acquisitions by foreign banks and several mergers among private domestic banks. Further

mergers and acquisitions are expected to follow in the near future.5

Table 3 below presents data on a group of privately owned commercial banks in Turkey

as of the end of September 2001. The first group, consisting of banks with assets in the USD 5-

10 billion range, is what can be considered as the "dominant" group of banks with market

power. The total assets of the 5 banks in this group make up about 64% of the assets of all

privately owned commercial banks, domestic as well as foreign.6 Most of the rest of the banks

listed in Table 3 are considerably smaller in size than the first (dominant) group. This group can

be considered as the "fringe" without market power. The total assets of the 20 banks listed in

Table 3 make up about 95% of the assets of all privately owned commercial banks, and the same

ratio for deposits and loans are even higher.7

TABLE 3 ABOUT HERE

3. A Stylized Model of the Banking Industry

In the Industrial Organization approach we adopt here, a bank is viewed as a firm that uses labor

and physical capital as inputs to produce different financial services for depositors and

5 An earlier version of this paper predicted a Turkish banking sector with about 20 private commercial banks (see Mumcu-Serdar et al., 2001). Table 1 reveals that the number of private commercial banks already fell to 24 by the end of September 2001. 6 In fact, after September 2001, Garanti, the fourth largest private bank in Table 3, merged with Osmanlı, the sixth largest private bank according to Table 3, thereby becoming the largest private commercial bank in Turkey. With this merger, the share of the first five largest private commercial banks in total sector assets now reached 69.8%. 7 In these calculations, foreign owned commercial banks have been included among the privately owned commercial banks. The banks under SDIF management have not been included in this category.

10

borrowers.8 The banking activity is thus viewed as the production of deposit and loan services.

In addition, we model the banking industry as an imperfectly competitive environment in which

banks act as independent entities. As an approximation of the structure of the Turkish banking

industry we will assume that there are few dominant banks together with a competitive fringe.

The economy is assumed to be open so that the banking industry can borrow and lend freely in

the world financial markets.

Specifically, we assume there to be a total of N domestic banks operating in an

oligopolistic industry. There are assumed to be n banks in the dominant group while the

competitive fringe consists of m banks, with n + m = N. Each bank is characterized by a cost

function,

( ), ,i i i i

c D L F= ,

where Di is the deposits collected and Li is the loans extended by bank i, and Fi is bank i's net

position in the world financial market. The balance sheet of a typical bank will therefore look as

displayed below:

Assets Liabilities

Loans (L) Deposits (D)

Reserves (R) Net Foreign Position (F)

Note that the reserves for a bank i is given by

i i i iR D L F= − + .

Cash reserves, Ci, and the net position on the interbank market, Mi, will make up bank i's

reserves Ri.

8 For various models of banking industry structure and bank behavior, see Freixas and Rochet, 1996.

11

The cash reserves Ci bear no interest, and therefore will be set at the minimum level

required by the Central Bank. Letting α be the reserve requirement ratio on domestic deposits

and β be the reserve requirement ratio on foreign credits set by the Central Bank, we will have

, for all i i i

C D F iα β= + .

The banks collect the deposits from households and borrow from international financial

markets, and channel these funds to firms to finance their investment needs. The third actor in

the real sector is the government. The government finances its deficit G either by issuing

securities ∆B or by issuing high-powered money ∆M0. Assuming that households do not hold

cash, the high-powered money is only used to finance banks' compulsory cash reserves held in

their accounts at the Central Bank. That is,

∑∑∑===

+==N

i

i

N

i

i

N

i

i FDCM111

0 βα .

We will assume that banks are differentiated in terms of their perceived riskiness in the domestic

financial market as well as in the foreign financial markets. Each bank i faces an upward sloping

inverse supply function of savings ( )D

ir S , where S B D= + , with B being the outstanding

government securities and ∑=

=N

i

iDD1

. Banks' different perceived financial strengths are

assumed to be summarized by parameters 1iλ ≥ such that ( ) ( )D D

i ir S r Sλ= , where ( )Dr S is

the upward sloping inverse supply function of savings faced by the least risky bank (i.e. the

bank i* with * 1iλ = ). Each bank i is assumed to face a downward sloping inverse demand

function of loans ( )Lr L , where ∑

=

=N

i

iLL1

. The cost of borrowing from the international

financial market for bank i is F F

i ir rµ= , where 1

iµ ≥ is bank i's perceived strength in the

12

international financial market and rF is the risk free foreign interest rate. Finally, banks can

borrow and lend in the interbank market at rate r.

3.1. The Behavior of the Dominant Group and the Competitive Fringe

The choice variables of the banks are deposits Di, loans Li, and net foreign position Fi. Taking

into account the operational costs, the profit of bank i will be given by

( ) ( ) ( ) ( ), , , ,L D F

i i i i i i i i i i i i i iD L F r L L rM r S D r F c D L Fπ λ µ= + − − − , (1)

where Mi, the net position of the bank i in the interbank market, is

( ) ( )1 1i i i iM D F Lα β= − + − − . (2)

Substituting (2) in (1) yields

( ) ( ) ( ) ( ) ( ) ( ), , 1 1 , ,L D F

i i i i i i i i i i i i iD L F r L r L r r S D r r F c D L Fπ α λ β µ⎡ ⎤ ⎡ ⎤ ⎡ ⎤= − + − − + − − −⎣ ⎦ ⎣ ⎦ ⎣ ⎦ .(3)

Thus, the banks' profits are the sum of the intermediation margins on loans, deposits, and the net

foreign position net of operational costs.

The banks in the competitive fringe are price takers in all three markets. That is, in

addition to the risk free foreign rate rF, they take the loan rate r

L and the deposit rate rD

determined by the dominant group as given as well. The first order conditions for the banks in

the competitive fringe are

( )1 0f

Di ii

i i

cr r

D D

π α λ∂ ∂= − − − =

∂ ∂, (4)

0=∂∂

−−=∂∂

i

iL

i

fi

L

crr

L

π, (5)

and

( )1 0f

Fi ii

i i

cr r

F F

π β µ∂ ∂= − − − =

∂ ∂. (6)

13

Let ( ), , ,f L D F

i i iL r r r r , ( ), , ,f L D F

i i iD r r r r , and ( ), , ,f L D F

i i iF r r r r be the loan supply, deposit

demand, and net foreign position of the banks in the competitive fringe, respectively, found by

solving (4), (5), and (6) simultaneously. Define

( ) ( )1

, , , , , ,m

f L D F f L D F

i i i i i

i

L r r r r L r r r r=

≡∑ , (7)

( ) ( )1

, , , , , ,m

f L D F f L D F

i i i i i

i

D r r r r D r r r r=

≡∑ , (8)

and

( ) ( )1

, , , , , ,m

f L D F f L D F

i i i i i

i

F r r r r F r r r r=

≡∑ (9)

as the total loan supply, the total deposit demand, and the total net foreign position of the fringe,

respectively. The residual demand for loans function that the dominant group will face will be

( ) ( )r L L fL r L r L= − , where ( )L

L r is the total demand for loans (inverse of ( )Lr L ). Similarly,

the residual supply of deposits function will be ( ) ( ), , , ,r D D fS r S r D Bλ µ λ µ= − − , where

( ), ,r DS r λ µ is the total supply of savings, with ( )1, ,

Nλ λ λ= ⋅⋅⋅ and ( )1, ,

Nµ µ µ= ⋅⋅⋅ . The banks

in the dominant group are assumed to engage in quantity (Cournot) competition in determining

their supply of loans and demand for deposits. Given the behavior of the competitive fringe, the

first order conditions summarizing the dominant group banks' maximizing behavior are

( ) ( ) ( )1

1, ,1 , , 0

r Ddr Di i

i i

i i i

S r cr D S r

D D D

λ µπ α λ λ µ−

−⎛ ⎞∂∂ ∂⎜ ⎟= − − + − =⎜ ⎟∂ ∂ ∂⎝ ⎠

, (10)

( ) ( )1

1

0r Ld

r Li ii

i i i

L r cL L r r

L L L

π−

−∂∂ ∂= + − − =

∂ ∂ ∂, (11)

and

( )1 0d

Fi ii

i i

cr r

F F

π β µ∂ ∂= − − − =

∂ ∂. (12)

14

Let ( ), , ,d L D F

i i iL r r r r , ( ), , ,d L D F

i i iD r r r r , ( ), , ,d L D F

i i iF r r r r be the loan supply, deposit

demand, and net foreign position of the banks in the dominant group, respectively, found by

solving (10), (11), and (12) simultaneously. Define

( ) ( )1

, , , , , ,n

d L D F d L D F

i i i i i

i

L r r r r L r r r r=

≡∑ , (13)

( ) ( )1

, , , , , ,n

d L D F d L D F

i i i i i

i

D r r r r D r r r r=

≡∑ , (14)

and

( ) ( )1

, , , , , ,n

d L D F d L D F

i i i i i

i

F r r r r F r r r r=

≡∑ (15)

as the total loan supply, the total deposit demand, and the total net foreign position of the

dominant group, respectively.

3.2. Equilibrium

Summing condition (2) over all banks will give the market clearing condition in the interbank

market as

( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )1 1d f d f d fD D F F L Lα β⎡ ⎤ ⎡ ⎤− ⋅ + ⋅ + − ⋅ + ⋅ = ⋅ + ⋅⎣ ⎦ ⎣ ⎦ . (16)

The market clearing conditions in the loan market and the deposit market will be given by

( ) ( ) ( )L d fL r L L= ⋅ + ⋅ , (17)

and

( ) ( ) ( ), ,D d fS r B D Dλ µ = + ⋅ + ⋅ , (18)

respectively. Note that all of the banks are assumed to be price takers in the foreign credit

markets. Solving (16), (17), and (18) simultaneously will give the non-cooperative equilibrium

in the banking industry for given N, λi, µi, α, β, and rF. Some of the comparative statics

exercises that can be carried out thus involve changes in N, n, and m (impact of mergers);

15

changes in α and β (impact of reserve requirement ratios); changes in λi (impact of deposit

insurance); µi (impact of country's international credit rating, or acquisitions by foreign banks);

and rF (impact of global economic conditions).

4. Empirical Application: Mergers and Acquisitions

We use special functional forms in simulating the model. As the inverse demand function for

loans and the inverse supply function of deposits we use

( )L

L Lr L A B L= − (19)

and

( )D

D Dr S A B S= + , (20)

respectively. As for the cost functions, we assume that the banks in the dominant group have a

cost advantage over those in the fringe.9 We assume that the costs of banks in the dominant

group are linear in deposits and loans, but quadratic in funds borrowed from abroad. The

functional specification adopted for the cost function of the banks in the dominant group is

( ) 2, ,2

fd d

i i i D i L i ic D L F K D L F

σσ σ= + + + (21)

The banks in the fringe are assumed to have decreasing returns to scale in production. Their cost

function is assumed to have the following form:

++= 2

2),,( i

Dfiii

f DKFLDcρ 22

22i

Fi

L FLρρ

+ (22)

The effect of the changes in the parameters of the model will be assessed mainly

according to their impact on welfare. We use the standard welfare measure of total surplus (the

sum of depositors' and creditees' surpluses and total profits of the banks) in quantifying welfare.

9 Evidence on economies of scale in banking the industry is mixed. Various studies show that only very small banks have a potential to achieve scale economies and that the average cost curve becomes quickly more or less flat for the larger firms (see Dermine, 1999, and references cited therein).

16

Mergers can take place either sequentially, which we will call sequential mergers,

whereby one bank is acquired at a time, or through block mergers , whereby a number of banks

are acquired together. The extent of consolidation will be much more under block acquisitions

than under sequential acquisitions. In the sequential case, the market power and profits of the

remaining banks are increased at each stage, making further acquisitions more costly for the

acquiring bank.

To the extent possible we use data from the Turkish banking industry. When data is not

available on a particular parameter, we concentrate on reasonable range of values and consider

the robustness of results to changes in the parameter values. The base case for parameters used

in most of empirical applications is summarized in Table 4 below. We used data from 1999, the

most recent data before the November 2000 and February 2001 crises.

TABLE 4 ABOUT HERE

We took the group of privately owned commercial banks listed in Table 3 above as

representing the completely privatized banking industry structure in our model. As already

indicated, the assets of these 20 banks make up about 95% of the assets of all privately owned

commercial banks in Turkey, and this ratio is even higher for deposits and loans. In the

dominant group there are 5 banks, each with assets in the U.S.D 5-10 billion range. The total

assets of these banks make up about 64.2% of the assets of all privately owned commercial

banks. The rest of the banks listed in Table 3 are considerably smaller in size than the dominant

group. This group of 15 banks is taken as the fringe without market power. Thus, as the base

line values of n and m we take n0 = 5 and m0 = 15, respectively.

Using the 1999 figures, the reserve requirement of 6% is taken as the value for α, while

the liquidity requirement ratio for liabilities in foreign currency of 3% is taken as the value for β

17

In arriving at values for Kd and K

f we use the overhead costs to assets ratio of 6%

reported by Demirgüç-Kunt and Levine (1999).

Given the blanket deposit insurance in Turkey that every bank enjoys without

discrimination, in the base case the perceived financial strength (from the viewpoint of

depositors) of all banks are taken as equal, i.e. we take λd = λf = 1.

An average LIBOR rate for 1999 (expressed in terms of TL) is taken as the value for the

risk free cost of borrowing in international markets. The values for µd and µf are arrived at by

considering the difference in borrowing costs of representative banks from the dominant and the

fringe groups.10

4.1. Mergers among Fringe Banks

4.1.1. Will a fringe bank acquire another fringe bank?

The first question we consider is whether a fringe bank will have an incentive to acquire another

fringe bank. We first look at the case where the merged entity is still a fringe bank; the other

case we look at is where the merged entity becomes a dominant bank.

Table 5 exhibits, for the base case as well as with higher fixed costs for the fringe, the

change in the profits of the fringe banks as the size of the fringe varies. When a fringe bank

acquires another fringe bank and the merged entity still remains a fringe bank, the post-merger

profits of the fringe banks are higher, as observed in Table 5. However, a merger will actually

take place if and only if both the acquiring and the acquired firms have incentives to merge. The

maximum amount that a fringe bank will bid to acquire another fringe bank is the difference

between its pre- and post-merger profits. A targeted fringe bank will accept an offer to merge if

the bid it receives exceeds its pre-merger profits. Therefore, a merger will go through if and only

10 We thank Cevdet Akçay for providing us this information on the perceived riskiness of the privately owned Turkish commercial banks in international financial markets.

18

if the maximum amount that the acquiring firm will be willing to bid exceeds the minimum

amount the firm to be acquired is willing to accept.

TABLE 5 ABOUT HERE

We observe from Table 5 that in the base case there is no profitable acquisition of a

fringe bank by another fringe bank, regardless of whether the acquisitions are carried out

sequentially or in block. The increase in the profits of banks when the number of fringe banks is

reduced by one from 15 to 14 is less than the pre-merger profit of a single bank.

As the fixed costs of the fringe banks increase, block acquisitions do become profitable,

however. In the high fixed cost case presented in Table 5, 10 banks could profitably merge as a

block. When there are 15 fringe banks, the profit of each fringe bank is TL 3.75 trillion. With

10 fringe banks merged, there will be 6 banks altogether in the fringe. The minimum total post-

merger profit that will be sufficient to compensate each merging bank is TL 37.5 trillion. The

profit of the merged bank will be TL 38.58 trillion in this case, which exceeds TL 37.5 trillion.

We observe that sequential mergers are not profitable even in this case of high fixed costs.11

As mentioned above, a merger among fringe banks may lead to their acquiring enough

resources to become one of the dominant banks. The comparative statics exercise that is

involved here requires checking whether there exist incentives for the fringe banks to become

dominant banks. Table 6 below shows that the profits of a fringe bank will increase when it is

converted to a dominant firm, as to be expected. If the increase in profits exceeds the cost of

acquiring the requisite technology to operate as a dominant bank, a single fringe bank will

indeed choose to become a dominant bank.12 For example, the profit of a fringe bank in an

industry with m = 15 and n = 5 is TL 33.7 trillion, while the profit of a dominant bank in an

11 As fixed costs do not affect the marginal decisions of banks, ∆πf

is the same for all levels of fixed costs. However, fixed costs affect the level of profits and thus the feasibility of mergers. 12 The assumption here is that there are no other extra-economic barriers that would prevent a fringe bank from acquiring the requisite technology to become a dominant bank.

19

industry with m = 14 and n = 6 is TL 82.6 trillion. Therefore, if the cost of acquiring the

requisite technology to operate as a dominant bank is less than TL 48.9 trillion, a fringe bank

will choose to become a dominant bank.

TABLE 6 ABOUT HERE

Also observe from Table 6 that two fringe banks could profitably merge to become the

sixth dominant bank in an industry structure with m = 13 and n = 6. The profit of a dominant

bank in the industry structure with m = 13 and n = 6 is TL 96.1 trillion. This exceeds TL 67.4

trillion, which is the total pre-merger profits of the two fringe firms, by 28.7 trillion TL. Note

that the profit advantage of becoming a dominant bank decreases in this case, making mergers

among fringe firms less likely. We observe from Table 6 that it would not be profitable for three

fringe banks to merge and then convert into a dominant bank.

4.1.2. Should Fringe Banks Merge?

We now turn to evaluating the desirability of mergers to see whether they conflict with

the incentives of banks to merge. As can be seen from Figure 1, for the base case of parameters

we consider, merging firms in the fringe may in fact lead to a decrease in total surplus as long as

the size of the dominant group remains the same (the solid line in Figure 1). Recall from above

that there were no incentives on the part of fringe firms to merge in this case either.

FIGURE 1 ABOUT HERE

On the other hand, if the fixed costs of the fringe banks were higher, it would be welfare

improving to reduce the number of banks in the fringe through consolidations among

themselves. In Figure 1 the dashed curve displays such a case, where the total surplus

maximizing number of firms in the fringe (when n = 5) is 12, indicating desirability of mergers

20

with respect to the base case where m is 15. Though we would expect total surplus to increase

with more competition, increasing m might not lead to higher welfare because of the fixed costs

of operation. If fixed costs are sufficiently small, then increasing the number of banks in the

fringe will always be welfare increasing. This is due to decreasing returns to scale in production

for the fringe. Recall from above that, though sequential mergers were not profitable, block

mergers did turn out to be profitable in this case. Note, however, that the reduction in the size of

the fringe (to a total of 6 banks) leads to a smaller fringe than the optimal size identified in

Figure 1.

Figure 2 displays the impact of changing the size of both the dominant and the fringe

groups on total surplus. We observe that merging a number of fringe banks and turning them

into a dominant bank will not necessarily increase welfare. For example, combining two fringe

banks (thus reducing m to 13) and turning them into one dominant bank (thus making n = 6) will

reduce welfare. This is despite the fact that competition among the dominant group has

increased with the addition of one firm, and the improved variable cost when the fringe firms are

merged and become a dominant bank. The high fixed cost of operating a dominant firm may

exceed gains from increased competition and improved production efficiency.

FIGURE 2 ABOUT HERE

But if cost structure is different, e.g. when fixed costs are lower, the welfare result

displayed in Figure 2 may be reversed. Figure 3 displays a case where it will desirable to merge

the banks in the fringe and turn them into a dominant bank.

FIGURE 3 ABOUT HERE

Note also from Figure 3 that whether it is desirable to merge the fringe banks or not will

depend on how many dominant firms there are. If there are four banks in the dominant group it

21

will not increase welfare to reduce the number of fringe banks. If, however, n is larger, then

mergers in the fringe become desirable.

4.2. Mergers among Dominant Banks

4.2.1. Will a dominant bank acquire another dominant bank?

Table 7 presents the change in the profits of the-dominant banks as the size of the dominant

group varies in the base case. Whether sequential or block mergers among dominant banks will

be profitable depends very much on their fixed costs of operation. In the base case where fixed

costs are relatively low, the increase in profits when competition is lessened with a sequential

merger will not exceed the pre-merger profits of a targeted dominant bank. Block mergers will,

however, be feasible: four dominant banks can profitably merge at once, leading to a two-bank

dominant group. On the other hand, when fixed costs are high, profits in the pre-merger stage

are low, making mergers more likely. In the case presented in Table 7 (with Kd = 0.25), both

sequential and block mergers are feasible.

TABLE 7 ABOUT HERE

4.2.2. Should dominant banks merge?

Whether dominant banks should merge or not depends on the fixed costs in the dominant group

as well as on the size of the fringe. We can use Figure 2 to consider the impact of changing the

size of the dominant group on total surplus. Though there is no incentive for the banks

themselves to merge in this case, it turns out that shrinking the size of the dominant group to

4n = would increase total surplus when m = 15. Further reducing the size of the dominant

group to n = 3 would reduce the total surplus, however. Reducing the dominant group by one

bank saves on the fixed costs at the expense of reduced competition; the competition effect takes

22

over when there is further reduction in the number of banks and the negative impact on total

surplus exceeds the savings from fixed costs. If fixed costs in the dominant group were higher,

further reducing the size of the dominant group would have been warranted.

One other factor that allows welfare improving reduction in the size of the dominant

group is the presence of a sizeable fringe with 15 banks. If the competitive fringe were smaller,

say, with m = 6, increasing the size of the dominant group would have increased monotonically

the total surplus.

4.3. Mergers Involving Both Dominant and Fringe Banks

4.3.1. Will a dominant bank acquire a fringe bank?

Consider now the case where fringe banks are bought out by a dominant bank in a sequential

manner (so m decreases while n remains constant). Table 8 shows that, in the base case, there

turns out to be no incentive for such acquisitions of fringe banks by a dominant bank. For block

mergers to be profitable, one dominant bank will have to acquire 13 fringe banks at once, which

indicates that this kind of mergers will be difficult to realize. When the fixed costs of the fringe

banks are higher, then both sequential as well as block acquisitions of fringe banks by a

dominant bank are feasible. For example, when Kf = 0.05, it will be desirable for a dominant

bank to acquire fringe banks sequentially until the size of the fringe falls to 12. It would also be

profitable to acquire the whole fringe as a block (seeTable 8).

TABLE 8 ABOUT HERE

4.3.2. Should a dominant bank acquire a fringe bank?

Welfare analysis in this case is similar to evaluating the desirability of mergers among fringe

banks while keeping the size of the dominant group intact. Figure 1 demonstrates, for the base

23

case, that the acquisition of fringe banks by dominant banks may lead to a decrease in total

surplus (the solid line in Figure 1). On the other hand, if the fixed costs in the fringe banks are

higher the acquisition of fringe banks by dominant banks will be welfare improving (the dashed

curve in Figure 1).

4.4. Changes in Policy Variables

4.4.1. Reserve Requirement Ratio

Figure 4 displays the impact of changing the reserve requirement ratio from 6% to 4%. This

policy change does not have any impact on the desirability of mergers in the fringe. Decreasing

α has the following impact on the equilibrium outcomes in this example: r decreases, rL

increases, rD decreases, the profits of both the dominant and the fringe banks decrease, and total

welfare increases. A decrease in a amounts to cost reduction which favors the depositors and

creditees.

FIGURE 4 ABOUT HERE

4.4.2. Deposit Insurance

Figure 5 displays the impact of changes in the perceived financial strength of the banks. In the

base case we consider, the dominant and the fringe banks are assumed to have the same

financial strength from the view point of the depositors. This is justified in the presence of

blanket deposit insurance extended to depositors of all banks without discrimination, which has

been the case in Turkey since 1994. Any change in the deposit insurance system that will lead to

differences in the perceived financial strength of banks will be expected to change equilibrium

outcomes. In the example exhibited in Figure 5, an increase in the perceived riskiness of the

24

fringe banks does not have an impact on the desirability of mergers among the fringe firms.

Increasing λf has the following impact on the equilibrium outcomes in this example: r increases,

rD decreases, rL increases, the profits of the dominant banks increase while those of the fringe

banks decrease, and total surplus decreases. Though the change in total surplus is rather small in

the case exhibited in Figure 5, the increase in the market share of dominant banks in deposits

may be considerable, leading to driving the fringe banks out of the sector.

FIGURE 5 ABOUT HERE

4.5. Acquisition by Foreign Banks

Consider a scenario in which foreign banks establish considerable presence in the Turkish

banking industry through acquisitions. If the dominant group is now dominated by foreign

banks, one implication of this will be the ease of access to international financial markets. We

can capture this change by considering a reduction in µd. Figure 6 displays the impact of such a

change on total welfare. In the example exhibited in Figure 6, a decrease in the perceived

riskiness in the international markets of the dominant banks does not have any impact on the

desirability of mergers among the fringe. Decreasing µd has the following impact on the

equilibrium outcomes in this example: r decreases, rD decreases, rL decreases, the market share

of dominant banks in deposits decreases, the profits of the dominant banks increase and those of

the fringe banks decrease, and total welfare increases.

FIGURE 6 ABOUT HERE

25

4.6. Changes in Global Economic Conditions

In our model the impact of the economic conditions external to the country is summarized by rF.

In Figure 7 below we look at the impact of changes in this parameter. If the conditions in the

world economy worsen so that rF increases, the equilibrium outcomes move in the following

directions in the example exhibited in Figure 7: r, rD, and rL all increase; the profits of both the

dominant and fringe banks decrease; and total welfare also decreases. Note that mergers among

fringe banks now become desirable.

FIGURE 7 ABOUT HERE

5. Concluding Remarks

We know from the Industrial Organization literature that a merger in an imperfectly competitive

industry involves two main effects. One is a decrease in competition due to the reduced number

of players in the market. The only countervailing effect that can balance this negative impact is

the possible efficiency gains that may be realized at the end of the merger. This will more likely

be the case in industries where production exhibits increasing returns to scale. In such a case the

merger will allow taking advantage of economies of scale and result in more efficient overall

production in the industry. Whether the increase in welfare due to efficiency gains or the

adverse effect of reduced competition will outweigh the other will in general depend on the

specifics of the industry.

In a simple but rigorous model, we studied a series of scenarios and assessed the likely

impact of mergers and acquisitions in the Turkish banking industry. Our exploration clearly

shows the importance of quantifying empirically the cost structures and other related parameters

for accurate assessment of the pros and cons of mergers in the banking industry. Depending on

parameter values, especially the fixed costs of operating a bank, we identified cases where

26

mergers among fringe banks as well as the dominant banks turn out to be desirable from the

welfare point of view.

Note that the welfare measure we used to evaluate changes in banking industry structure

as well changes in policy variables was the standard total surplus measure that gave equal

weight to consumers’ (depositors’ and creditees’) surpluses and total industry profits. One can

certainly envision more general social welfare functions that give higher weight to depositers’

and creditees’ welfare, and our analyses can easily be repeated for more general social welfare

functions.13

Whether there will indeed be incentives to merge or not is a question related to, but

different than, the desirability of mergers. In cases where there is incentive to merge with

adverse welfare impacts, the competition authority may want to step in to block the proposed

merger. We identified such cases where dominant banks will have incentive to make bids to

acquire fringe banks that will be acceptable to the targeted banks. We also showed that whether

the mergers proceed sequentially, i.e. one firm is acquired first and then another, or a number of

firms are acquired in block makes a difference for the desirability of mergers. As far as

incentives to merge are concerned, the extent of consolidation will be greater under block

acquisitions than under sequential acquisitions. In the sequential case, the market power and

profits of the remaining banks are increased at each stage, making further acquisitions more

costly for the acquiring bank.

When mergers involving only the fringe firms are considered, we found that as the fixed

costs of the fringe banks increase block acquisitions did become profitable. High fixed costs in

this case also made decreasing the number of fringe firms preferable from the welfare point of

view. The extent of consolidation that would come about could, however, be more than the

optimal size identified. In the case of a number of fringe banks merging and becoming a

13 If there is concern regarding the stability of the banking system, giving less weight to profits may lead, in general, to more frequent exit and entry in the sector than desirable. In turn, this may compromise systemic stability of the sector. In addition to being the standard assumption in the Industrial Organization literature, total surplus reflects a neutral stance regarding this point.

27

dominant bank, total surplus will in certain cases decrease. This is despite the fact that

competition among the dominant group has increased with the additional banks joining the

group, and the improvement in variable cost when the fringe banks become larger. The high

fixed cost of operating a large bank may exceed gains from increased competition and improved

production efficiency.

Whether sequential or block mergers among dominant banks will be profitable depends

very much on their fixed costs of operation. As the fixed costs of operating a large bank rise,

block mergers as well as sequential mergers may become feasible. Whether dominant banks

should merge or not depends on the fixed costs in the dominant group as well as on the size of

the fringe.

When the fixed costs of the fringe banks are high, then both sequential as well as block

acquisitions of fringe banks by a dominant bank become feasible. Welfare analysis in this case

demonstrates that the acquisition of fringe banks by dominant banks may increase or decrease

total surplus depending on the fixed costs of the fringe banks.

Changes in the reserve requirement ratio, the extent of deposit insurance, and improved

perception of riskiness in international financial markets do affect the equilibrium outcomes

obtained. On the other hand, incentives to merge do not seem to be significantly affected by

changes in these variables. A dramatic effect of removing the blanket deposit insurance may be

the forcing of the fringe banks out of the sector, which will reduce total surplus. We also

demonstrated how worsening conditions in the world economy may render consolidation of the

fringe desirable from the welfare point of view.

A number of lessons can be drawn from our analysis regarding mergers in the banking

sector. As for whether mergers will occur or not, the nature of the merged entity is of critical

importance. If two fringe banks merge to become a dominant bank, this indicates not only an

improvement in the cost structure but also enhanced market power. Thus even in cases where

there is no incentive for fringe firms to merge to remain a fringe firm, they will merge if through

28

merging they can exercise market power. The welfare consequences of such mergers have to be

carefully analyzed, paying due attention to the relative size of the fringe and the dominant group

and the cost structure in the banking industry before and after mergers. Consolidations among

both the fringe and the dominant banks may turn out to be excessive, and thus their extent has to

be carefully assessed by competition authorities. Though block mergers, where a group of banks

are acquired by another bank, are not necessarily welfare reducing, this type of mergers may end

up stifling competition too much. Sequential mergers, where only one bank can be acquired at a

time, are more difficult; thus, allowing only this type of mergers will be a safety check against

excessive mergers.

Finally it is to be noted that mergers among dominant as well as fringe banks will be

impeded by a prisoners’ dilemma type interaction among each other. Consolidation in the sector

will benefit all banks through reduced competition. But those that are not actually involved in

mergers will not bear merging costs. For example, Table 1 reveals that if 10 fringe banks merge

to reduce the size of the fringe from 15 to 6, those that do not take part in the merger will enjoy

a profit increase of TL 34.83 trillion. Compared to its pre-merger profit, each of the 10 merging

fringe bank will be able to derive much less benefit from the merger, since the total profits of the

merged entity will not be more than the profit of fringe firms that remained out of the merger.

Thus, each bank favors mergers in the sector but does not want itself to be a part of mergers.

Unless mergers lead to dramatic improvement in the cost structure and lead to significant

increase in market share, which in the case of fringe banks can only happen if the merged entity

becomes one of the dominant banks, the prisoners’ dilemma identified will render consolidation

in the sector highly unlikely.

29

Acknowledgements

We gratefully acknowledge the contribution of Mehmet Aras Orhan for an earlier version of this

paper. We also thank two anonymous referees for helpful comments. Göze Aslankara provided

assistance with data for the current version, which we would like to acknowledge. The paper

draws from a project, titled “The Restructuring of the Turkish Banking Industry and Its Impacts

on Competition”, funded by Boğaziçi University Research Fund, Project No. 00C103.

References

Banks Association of Turkey, 2000. Turkish Banking System - Quarterly Statistics by the

Groups 1992 - September-2001. <http://www.tbb.org.tr/english> (December 30, 2001).

BRSA, 2001. ‘‘Towards a Sound Turkish Banking Sector’’, May 15.

<http://www.bddk.org.tr/english/mainpage/index_eng.htm> (December 30, 2001).

Demirgüç-Kunt, A., Levine, R., 1999. Bank-Based and Market-Based Financial Systems: Cross

Country Comparisons. Policy Reseach Working Paper No. 2143, The World Bank,

Washington D.C.

Dermine, J., 1999. The Economics of Bank Mergers in the European Union, A Review of the

Public Policy Issues. INSEAD Working Paper No. 99/35/FIN.

Freixas, X., Rochet, J.C., 1996. Microeconomics of Banking. MIT, Cambridge, Massachusetts.

Mumcu-Serdar, A., Zenginobuz, E. Ü., Orhan, M. A., 2001. The Likely Impact of Mergers and

Acquisitions on the Banking Sector: A Stylized Model and Simulations. Boğaziçi University

Research Papers, ISS/EC 2001-02.

30

Table 1: Turkish Banking Sector (September 30, 2001)

Number

of

Banks

Total

Assets

(USD

mn)

Share

in

Sector

(%)

Total

Loans

(USD

mn)

Share

in

sector

(%)

Total

deposits

(USD

mn)

Share

in

sector

(%)

TOTAL 68 110,763 100.0 28,642 100.0 68,796 100.0 Commercial

Banks

52 105,402 95.2 26,338 92.0 68,796 100.0

-State Owned 3 30,101 27.2 5,322 18.6 21,430 31.2 -Privately Owned 24 59,403 53.6 19,240 67.2 38,946 56.6 -5 largest banks 40,005 36.1 15,127 52.8 26,568 38.6 -Other 19,398 17.5 4,112 14.4 12,379 18.0 -Foreign Owned 16 2,933 2.6 681 2.4 978 1.4 -SDIF Managed 9 12,965 11.7 1,096 3.8 7,441 10.8 Development &

Investment Banks 16 5,330 4.8 2,304 8.0 0 0.0

-State Owned 3 4,347 3.9 1,868 6.5 -Privately Owned 10 881 0.79 400 1.4 -Foreign Owned 3 102 0.09 36 0.1

Sources: The Banks Association of Turkey, BRSA

31

Table 2: Turkish Banking Sector: Financial Indicators

In USD million (unless otherwise stated)

1980 1990 1994 1999 2000

Total Assets 20,785 58,171 52,552 133,533 155,237Total Credits 11,168 27,342 20,559 40,206 50,931Securities Portfolio 1,339 5,997 5,955 22,955 17,848Total Deposits 10,188 32,564 33,191 89,361 101,884Savings Deposits 4,288 19,343 24,190 58,807 64,352 -TL 4,288 11,914 8,612 24,701 26,628 -FX .. 7,429 15,578 34,106 37,724Non-deposit funding 1,289 11,760 9,019 22,934 29,435 -Foreign banks .. 3,460 2,675 12,073 16,284Networth+Profits 1,147 5,903 4,409 7,840 11,367 Total Assets/GNP (%) 28.6 38.2 40.3 71.7 76.9Total Credits/GNP (%) 15.4 17.9 15.8 21.6 25.2Securities Portfolio/GNP (%) 1.8 3.9 4.6 12.3 8.8Savings Deposits /GNP (%) 5.9 12.7 18.5 31.6 31.9State Bank Assets/Total Sector Assets (%)

44.1 44.6 39.6 34.9 34.2

Off-balance Sheet Operations/ Total Assets (%)

.. .. 49.5 103.5 100.8

Source: BRSA (2001)

32

Table 3: Private Commercial Banks Ranked by Asset Size (September 2001)*

Total

Assets

(USD

mn)

Share in

Sector

(%)

Total

Loans

(USD

mn)

Share in

sector

(%)

Total

deposits

(USD

mn)

Share in

sector

(%)

USD 5-10 billion

İş Bankası 9,491 8.6 3,247 11.3 7,084 10.3

Akbank 9,451 8.5 2,838 9.9 5,866 8.5

Yapı Kredi 8,638 7.8 3,425 12.0 5,839 8.5

Garanti 6,780 6.1 2,161 7.5 3,830 5.6

Pamukbank 5,644 5.1 3,456 12.1 3,949 5.7

Group Total 40,005 36.1

(64.2)

15,127 52.8

(75.9)

26,568 38.6

(66.5)

USD 1-5 billion

Osmanlı 3,536 3.2 507 1.8 1,485 2.2

Koçbank 3,127 2.8 988 3.5 2,086 3.0

Finansbank 2,326 2.1 610 2.1 1,151 1.7

Sümerbank 1,476 1.3 14 0.0 1,322 1.9

Dışbank 1,271 1.1 239 0.8 633 0.9

İmar Bankası 1,129 1.0 399 1.4 958 1.4

Toprakbank 1,123 1.0 203 0.7 1,015 1.5

TEB 1,056 1.0 267 0.9 712 1.0

Group Total 15,043

13.6

(24.1)

3,227 11.3

(16.2)

9,362 13.6

(23.4)

USD 0-1 billion Şekerbank 933 0.8 221 0.8 713 1.0

Denizbank 933 0.8 185 0.6 617 0.9

Alternatifbank 776 0.7 183 0.6 496 0.7

Tekstil Bank 595 0.5 69 0.2 397 0.6

Oyak Bank 355 0.3 104 0.4 259 0.4

Anadolubank 336 0.3 63 0.2 256 0.4

Tekfenbank 163 0.1 28 0.1 144 0.2

Group Total 4,092 3.7

(6.6)

854 3.0

(4.3)

2,881 4.2

(7.2)Sources: The Banks Association of Turkey, BRSA * As of September 30, 2001, there were 24 privately owned domestic banks and 16 foreign banks. Of these, the largest 20 in terms of total asset size are reported in this table. ** The figures in parentheses indicate the shares within all privately owned commercial banks, including the foreign owned.

33

Table 4: Base Model Parameter Values*

n0 = 5 AD = 0.01 Kd = 0.2 K

f = 0.03 λd= 1 rF = 0.03

m0 = 15 BD =0.01 σD =0.01 ρD = 0.04 λf =1 B = 9.7

α= 0.11 AL = 0.9 σL =0.02 ρL = 0.05 µd= 1.05

β= 0.11 BL =0.01 σF = 0.06 ρF = 0.15 µf= 1.1 * The TL values are in quadrillion for the relevant parameters (AD , AL , K

d , Kf , B).

34

Table 5: Mergers among Fringe Banks*

n = 5m πf

(Kf = 0.03) πf

(Kf= 0.06) ∆πf

1 135.57 105.57 2 112.70 82.70 22.85 3 96.85 66.85 15.85 4 85.11 55.11 11.74 5 75.97 45.97 9.14 6 68.58 38.58 7.39 7 62.43 32.43 6.15 8 57.19 27.19 5.24 9 52.63 22.63 4.56 10 48.62 18.62 4.01 11 45.05 15.05 3.57 12 41.82 11.82 3.23 13 38.90 8.90 2.92 14 36.22 6.22 2.68 15 33.75 3.75 2.47

*Profit figures ( πf ) are in TL trillion.

35

Table 6: Fringe Banks Merging into a Dominant Bank*

n = 5 n = 6

m πd πf πd πf

1 719.78 135.57 518.22 109.01 2 607,96 112.70 434.94 91.41

3 522.82 96.85 371.22 79.06

4 455.34 85.11 320.49 69.80

5 400,23 75.97 278.89 62.52

6 354.16 68.58 243.98 56.57

7 314.94 62.43 214.15 51.58

8 281.03 57.19 188.27 47.29

9 251.36 52.63 165.53 43.54

10 225.11 48.62 145.35 40.22

11 201.67 45.05 127.29 37.23

12 180.60 41.82 110.99 34.53

13 161.52 38.90 96.18 32.07

14 144.14 36.22 82.67 29.80

15 128.22 33.75 70.26 27.70

* Profit figures (πd , πf) are in TL trillion.

36

Table 7: Mergers Among Dominant Banks*

m = 15n πd (Kd

= 0.2) πd (Kd= 0.25) ∆πd

1 1593.22 1543.22

2 696.66 646.66 896.5

3 374.25 324.25 322.41

4 218.12 168.12 156.13

5 70.26 78.22 89.89 * Profit figures (πd) are in TL trillion.

37

Table 8: Dominant Firms Acquiring Fringe Firms*

n = 5

m πd πf (Kf

= 0.03) πf(K

f = 0.05) ∆πd

1 719.78 135.57 115.57 2 607.96 112.70 92.70 111.81

3 522.82 96.85 76.85 85.14

4 455.34 85.11 65.11 67.48

5 400.23 75.97 55.97 55.11

6 354.16 68.58 48.58 46.06

7 314.94 62.43 42.43 39.22

8 281.03 57.19 37.19 33.90

9 251.36 52.63 32.63 26.67

10 225.11 48.62 28.62 26.25

11 201.67 45.05 25.05 23.43

12 180.60 41.82 21.82 21.07

13 161.52 38.90 18.90 19.08

14 144.14 36.22 16.22 17.38

15 128.22 33.75 13.75 15.91

*Profit figures (πd , πf ) are in TL trillion.

38

Figure 1: Impact of Changing m keeping n constant

15

15,3

15,6

15,9

16,2

16,5

16,8

17,1

17,4

17,7

18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal S

urp

lus

n=5,Kf=0.03

n=5,Kf=0.06

39

Figure 2: Impact of Changing both m and n

16,2

16,4

16,6

16,8

17

17,2

17,4

17,6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal

Su

rplu

s

n=3

n=4

n=5

n=6

Figure 3: Impact of Changing Both n and m When Kd is Low

16,6

16,8

17

17,2

17,4

17,6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal S

urp

lus

n=4

n=5

n=6

40

Figure 4: Impact of Changing a

16,4

16,6

16,8

17

17,2

17,4

17,6

17,8

18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal S

urp

lus

n=5, alpha=0,06

n=5, alpha=0,04

41

Figure 5: Impact of Changing λf

16600

16700

16800

16900

17000

17100

17200

17300

17400

17500

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal S

urp

lus

n=5, lambdaf=1

n=5, lambdaf=1.05

42

Figure 6: Impact of Changing µd

15

15,3

15,6

15,9

16,2

16,5

16,8

17,1

17,4

17,7

18

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal S

urp

lus

n=5,Mud=1.05

n=5,Mud=1

43

Figure 7: Impact of Changing rF

15,3

15,5

15,7

15,9

16,1

16,3

16,5

16,7

16,9

17,1

17,3

17,5

17,7

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

m

To

tal S

urp

lus

n=5, rF=0.30

n=5, rF=0.38